The aim of this project is to study computer methods for designing conformal radiation therapy treatments for cancer. Automated delivery of radiation therapy permits the use of new conformal treatment strategies which uniquely meet the needs of individual patients but are too complex and sophisticated for manual delivery. Computer control removes limitations on numbers of beams, beam directions, and beam shapes, so that conventional methods of treatment planning using experience-guided trial- and-error search are totally inadequate. Efficient generation of clinically optimal computer-controlled treatment plans will require mathematical optimization techniques guided by an expert system advisor. Clinically realistic treatment objectives must include normal tissue complication probabilities and normal tissue dose-volume constraint limitations, both of which are non-linear non-analytic functions. Simulated annealing and downhill simplex and general mathematical methods for constrained optimization of such functions. With proper customizing, both can be applied to the creation of computer-controlled treatment plans. Different implementation methods, parameter values, and clinical objective/constraint functions will be studied for both methods using a variety of patient test cases. A variety of different computer-controlled delivery strategies are possible depending on the capabilities of the treatment machine. Dynamic rotation, conformal static beam, and non-coplanar beam techniques, with and without wedges, are practical with an automated multileaf collimator. Segmented conformal therapy and "field-in-field" techniques are deliverable using either an automated multileaf collimator or four independently-movable collimator jaws. The relative advantages of these different delivery techniques will be compared using mathematically optimized plans.